Allo transplant can result in cure, but the majority of pts relapse or die from treatment related complications. The objective of this proposal is the development of a more effective and less toxic transplant conditioning regimen. We plan to achieve this objective by completing the following specific aims : 1.To establish the maximally tolerated dose of intravenous busulfan (Busulfex) in combination with fludarabine as conditioning regimen for transplantation with in-vivo T-cell depletion: Bu is widely used in the conditioning for allogeneic transplantation. Its dose limiting toxicity is hepatic veno-occlusive disease (VOD). The risk for VOD is related to BU ADC, but is also influenced by other medications (e.g. cyclophosphamide, methotrexate) and procedures (e.g. T-cell depletion). We hypothesize that the maximum AUC will be increased when BU doses are targeted and other risk factors for VOD minimized. We plan a dose escalation study of IV BU (Busulfex) and fludarabine in combination with alemtuzumab GVHD prophylaxis. Individualized dosing to achieve a pre-defined target AUC, and thus predictable systemic exposure, will be based on pharmacokinetic analysis of a test dose. 2. To evaluate disease free and overall survival after dose escalated BuFluCampath conditioning. Prior studies indicate an inverse relationship between BU concentration and recurrence. We hypothesize that increased BU AUC will result in improvement in disease control. After the maximum AUC has been established, the efficacy will be studied in a prospective phase II study of patients with high risk AML and MDS. 3.To evaluate the effect of GSTA1 and GSTM1 polymorphisms and enzyme levels on kinetics and toxicity of BU. BU metabolism is mainly mediated by glutathione S-transferases A1 and M1. We hypothesize that variations in BU metabolism and toxicity are due to GST polymorphisms. We propose to evaluate polymorphisms and activity of GSTA1 and GSTM1 and to correlate them with BU metabolism and toxicity. 4. To study the relation between GST expression in leukemia cells. GST genotype and outcome of transplant. We hypothesize that resistance to busulfan is related to GST expression in leukemia cells, which in turn is partially related to GST genotype. We propose to study expression of GST's in leukemia cells and to correlate GST expression with genotype and with response to transplant. These studies are exploratory. [unreadable] [unreadable]